Beyond Isolated Behaviors: Hierarchical User Modeling for LLM Personalization
2026-06-01 • Computation and Language
Computation and Language
AI summaryⓘ
The authors propose a new way to personalize large language models (LLMs) by organizing user behaviors into three layered categories inspired by sociologist Pierre Bourdieu's theory. They group individual actions as practices, long-term habits as habitus, and common patterns among similar users as fields. Their method, called PHF Compass, uses this hierarchy to improve how models adapt to users without changing the model itself. Tests show this approach works better on different tasks and helps understand the learned behaviors more clearly.
Large Language ModelsPersonalizationPierre BourdieuTheory of PracticePracticeHabitusFieldBehavioral StructuresModel-AgnosticLaMP Benchmark
Authors
Liang Wang, Xinyi Mou, Xiaoyou Liu, Tiannan Wang, Yuqing Wang, Zhongyu Wei
Abstract
Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse domains, yet personalizing their outputs to individual users remains an open challenge. Existing approaches predominantly adopt a flat behavioral paradigm, aggregating user behaviors without an explicit account of how they are organized into deeper behavioral structures. In this work, we draw on Pierre Bourdieu's Theory of Practice to propose PHF (Practice-Habitus-Field), a sociologically grounded framework that reconceptualizes LLM personalization through three hierarchical levels: individual behaviors as practices, their temporal accumulation into stable dispositions as habitus, and shared regularities across similar users as fields. We instantiate PHF through $\mathrm{PHF}_{\text{Compass}}$, a lightweight and model-agnostic implementation based on a frozen LLM. Experiments on the Language Model Personalization (LaMP) benchmark demonstrate consistent improvements across diverse tasks, while further analyses validate the interpretability and extensibility of the learned behavioral structures.